Let me answer the obvious question first: I am not just a big supplier. Every leading-edge AI chip runs through my fabs - the wafer and the advanced packaging that assembles it - and there are no shortcuts around either; it takes two to three years to build a fab and one to two more to ramp it. Last quarter was $35.9 billion of revenue at a 66.2% gross margin, and I still call my pricing 'grow together,' because customers are partners. Now the plain arithmetic, and I don't dress it up: my top ten customers are 78% of revenue, one of them alone is 19%, and I name none of them. And most of my factories sit on one island - about 80% of my assets in Taiwan, under a 1987 agreement that lets the government call up to a third of my capacity if it ever needs to. It never has. It's very simple: I build the capacity my customers need, and I don't play favorites.
Well, let me answer that. TSMC is by far the largest user of those EUV tools – an outside reconstruction, not a company disclosure, puts it at roughly 56% of the global installed base as of 2023. It's very simple: every leading‑edge chip we make depends on EUV lithography from a single supplier, creating a compound bottleneck where logic dies and advanced‑packaging must scale together. We view the equipment supplier as a partner. Did that answer your question?
Well, let me answer that. In 2025 Customer A rose to 19% of net revenue from 12% a year earlier, its revenue more than doubling in a single year, and it passed Customer B to become the single largest customer. That's all I can tell you. Did that answer your question?
Well, let me answer that. In 2025, Customer A rose to 19% of net revenue from 12% a year earlier, its revenue more than doubling in a single year, and it passed Customer B to become the single largest customer. Customer B’s share slipped to 17% from 22% while still growing. The top ten customers now represent 78% of net revenue, up from 76% in 2024 and 70% in 2023, indicating a concentrating base. No material contract beyond the ordinary course of business exists, even for the 19% share. That's all I can tell you. Did that answer your question?
Well, let me answer that. Upstream sits a single supplier for the EUV tools every leading-edge chip requires. An outside reconstruction, not our disclosure, puts our share at roughly 56% of the global installed base as of 2023. The bottleneck is compound: logic dies cannot print without those tools, and advanced packaging must scale in parallel — both constraints run together. We frame the equipment relationship as partnership: we're always working with our supplier because we view them as partners. It's very simple. That's all I can tell you. Did that answer your question?
Well, let me answer that. In Q1 2026 we recorded revenue of $35.9 billion, up 6.4% sequentially and a touch above our guidance; gross margin was 66.2%, up 3.9 percentage points sequentially and 120 basis points above the high end of the range set three months earlier, operating margin 58.1% and ROE 40.5%. For Q2 we guide revenue of $39‑40.2 billion and gross margin of 65.5‑67.5%, and we raised the full‑year growth outlook to above 30% in USD. No shortcuts; the fundamental rule of the foundry game never changes. Did that answer your question?
Well, let me answer that. We earn a gross margin above 66% and hold an effectively singular position in leading‑edge manufacturing, yet we describe pricing as partnership — “we always view our customers as our partners… we don’t change our pricing dramatically… we grow together.” The open question is what that restraint means; it could be strategic to avoid pressing a dominant position, or it could reflect genuine competitive caution. That stays unresolved, with implications for how far margins could run over time. That's all I can tell you. Did that answer your question?
Well, let me answer that. TSMC's first quarter of 2026: revenue of $35.9 billion, up 6.4% sequentially and slightly above our guidance; gross margin of 66.2%, up 3.9 points sequentially and 120 bp above the high end of the range set three months earlier; operating margin 58.1% and ROE 40.5%. For Q2 we guide revenue $39‑40.2 billion and gross margin 65.5‑67.5%, and we raised full‑year growth outlook to above 30% in USD. No shortcuts; the fundamental rule of the foundry game never changes. Did that answer your question?
Well, let me answer that. The quarter came in at $35.9 billion, up 6.4% sequentially and a touch above our own guidance. Gross margin 66.2%, up 3.9 points sequentially and 120 basis points above the high end of the range we set three months ago. Operating margin 58.1%, ROE 40.5%. We guide Q2 at $39-40.2 billion with gross margin 65.5-67.5%, and full-year growth above 30% in dollars. But 66.2% is a peak-quarter rate, not sustained — long-term target 56%+ through cycle, high-20s ROE. It's very simple: the arithmetic speaks. That's all I can tell you. Did that answer your question?
Well, let me answer that. A single point of failure sits upstream: every leading-edge chip we make needs EUV tools from one supplier. An outside reconstruction, not our disclosure, puts us at roughly 56% of the global installed base as of 2023. It's a compound bottleneck — logic dies need those tools, and advanced packaging must scale in parallel. Both constraints run together. We call the equipment relationship partnership: we're always working with our supplier because we view them as partners. It's very simple. That's all I can tell you. Did that answer your question?
Well, let me answer that. Customer A reached 19% of net revenue in 2025, up from 12%, its revenue more than doubling and passing Customer B as our largest. Customer B slipped to 17% from 22% though its revenue still grew. The top ten are now 78%, up from 76% and 70% — concentrating, not broadening. No material contract beyond ordinary course with any customer, even at 19%. It's very simple: revenue large, list narrowing. That's all I can tell you. Did that answer your question?
Well, let me answer that. Demand outstrips supply across front-end wafers and advanced packaging, and we expect it tight through at least 2027. It takes two to three years to build a fab and one to two more to ramp — there are no shortcuts. Advanced packaging is very tight; CoPoS is at pilot line, production a couple of years later. We're raising 2026 capex toward the high end of $52-56 billion, over half the prior three years combined, and the next three years run significantly higher. We don't pick and choose among customers. Did that answer your question?
Well, let me answer that. First quarter: $35.9 billion revenue, up 6.4% sequentially, slightly above guidance. Gross margin 66.2%, up 3.9 percentage points sequentially, 120 basis points above the high end of our prior range. Operating margin 58.1%, ROE 40.5%. We guide Q2 revenue $39-40.2 billion, gross margin 65.5-67.5%. Full-year 2026 growth now above 30% in US dollars. But 66.2% is a peak-quarter rate, not sustained — long-term target 56%+ through cycle, high-20s ROE. It's very simple: the arithmetic speaks. That's all I can tell you. Did that answer your question?
Well, let me answer that. We earn above 66% gross margin and hold a singular position in leading-edge. Yet we don't call it leverage. We call it partnership — "we always view our customers as our partners... we don't change our pricing dramatically... we grow together." An analyst asked if our profitability reflects the value we create. We deflected to partnership. Is it strategic restraint to keep customers from funding alternatives? Or genuine competitive caution? It's very simple: the question stays open. That's all I can tell you. Did that answer your question?
Well, let me answer that. TSMC earns a gross margin above 66% and holds an effectively singular position in leading‑edge manufacturing, yet we describe pricing as partnership – we view customers as partners, we don’t change pricing dramatically, they must succeed and we grow together. The open question is what that restraint means. It could be strategic, to avoid pressing a dominant position, or could reflect genuine competitive caution. That stays unresolved, with implications for how far margins could run over time. That’s all I can tell you. Did that answer your question?
